A clustered data state is the condition of data that has been through a clustering process.1

In the SWEET ontology the class is a subclass of data state, restricted to data that had a process of the clustering type.1 SWEET treats cluster as equivalent to clustering and as a kind of data analysis service.2

Clustering in remote sensing divides pixels into groups by natural groupings in their spectral values. A NASA report describes unsupervised classification as dividing a large number of pixels into classes based on natural groupings present in the image, usually by clustering methods.3 The Esri Iso Cluster tool separates all cells into a user-specified number of groups in the multidimensional space of the input bands.4

Retained sources do not establish how a product records this state in metadata, nor which clustering method it implies. The ontology gives the restriction but no prose definition.1

Footnotes

  1. ESIP, “SWEET ontology: stateDataProcessing module (Turtle source)”. https://raw.githubusercontent.com/ESIPFed/sweet/master/src/stateDataProcessing.ttl ↩ ↩2 ↩3

  2. ESIP, “SWEET ontology: reprDataServiceAnalysis module (Turtle source)”. https://raw.githubusercontent.com/ESIPFed/sweet/master/src/reprDataServiceAnalysis.ttl ↩

  3. NASA Technical Reports Server, “Cluster Method Analysis of K.S.C. Image”. https://ntrs.nasa.gov/api/citations/20010000468/downloads/20010000468.pdf ↩

  4. Esri, “How Iso Cluster works”. https://doc.esri.com/en/arcgis-pro/latest/tool-reference/spatial-analyst/how-iso-cluster-works.html ↩